684 Prof. 0. W. Richardson on the 



These results are particularly instructive. Iu the ex- 

 periments with + 40 volts the temperature was very constant, 

 the extreme variation being 4° C. There is, nevertheless, a 

 considerable difference between the emission at a given 

 pressure on the outward and return journeys. This is owing 

 to a decay of the emission with time. This can be eliminated 

 witn sufficient accuracy by taking the mean of the two 

 curves. In the experiments with + 160 volts there was a 

 rapid fall of temperature, the extreme variation being 63° C, 

 which caused a greater relative difference between the out- 

 ward and return curves. This again should be eliminated 

 along with the time changes by taking the average. When 

 this is done it is seen that the general march of the curve is 

 the same as that of the average curve with +40 volts. This 

 shows that the effect on the mean curve of a steady drift in 

 the temperature is practically negligible ; and it is important 

 because it was found impossible to keep the temperature 

 constant owing to lack of steadiness in the supply of current. 

 Both of the above cases, however, are extreme samples ; in 

 nearly all the other experiments there was a drift of between 

 ten and twenty degrees. In the experiments with almost 

 all the other salts the conditions in the series of experiments 

 with +40 and +160 volts were practically identical, except 

 for the difference in voltage, and in nearly every case the 

 curves were indistinguishable, except for a difference in the 

 absolute value of the currents. In considering them it will, 

 therefore, only be necessary to refer to a single voltage, 

 except where there are special reasons for doing otherwise. 



Returning to the mean curves for sodium sulphate, it will 

 at once be observed that they are quite different from those 

 obtained by the strip method. In the first place the emission 

 is not sensitive to small changes in the pressure at low 

 pressures. A change from 0'004 mm. to 0*15 mm. pressure 

 only increases the emission from 28 to 31 divisions, whereas 

 in an experiment by the strip method at a similar tempera- 

 ture an increase of pressure from 016 mm. to 0*15 mm. 

 more than doubled the emission. Some typical cases are 

 exhibited graphically in fig. 1. Curves 1 and 2 in this 

 diagram were both obtained by the strip method at 730° C. 

 In the case of curve No. 1, which gave a maximum ioniza- 

 tion at a pressure of 20 mm., the specimen was comparatively 

 fresh, whereas in the case of curve No. 2 it had been heated 

 for a long time and at high temperatures. Although these 

 curves differ considerably from one another, they both differ 

 from No. 3 in being much more sensitive to changes of 

 pressure at the lowest pressures. 



